08:01 mblenc: hi, i have an issue with my wayland compositor and an external monitor ive hooked up. the model number is WJ1980PI, and it is hooked up via a hdmi-to-vga adapter, as the monitor only have a soldered vga-in
08:02 mblenc: the monitor works fine under windows, but under wayland and X it constantly resets when anything bright (i.e. the google.com homepage) is placed on it
08:02 mblenc: this issue persists across multiple wayland compositors (ive tried dwl and sway), and nothing shows up under dmesg, so im a little confused
08:03 mblenc: where can i find additional logs for this kind of issue? and if this is not the correct place to ask for support, where should i direct my query to?
08:10 MrCooper: sounds like a KMS driver issue
08:22 mblenc: how can i debug that?
08:24 mblenc: for a video showcasing the error: https://mblenczewski.com/pastebin/monitor-reset.mp4
08:25 emersion: check dmesg maybe
08:26 mblenc: dmesg shows no errors relating to KMS, or otherwise, when the monitor resets
08:27 orowith2os[m]: MrCooper: if it were a KMS driver issue, it wouldn't happen on X though, would it?
08:27 orowith2os[m]: IIRC X still does some weird framebuffer f*ckery
08:32 MrCooper: not sure what you mean by that, Xorg has been using KMS since before Wayland was a thing
08:39 orowith2os[m]: I need to look into it a bit more, or have someone explain it to me, bleh
08:40 orowith2os[m]: my understanding thus far has been that X liked to go directly to a framebuffer, something something root userspace
08:40 orowith2os[m]: and kms drivers have been available, but were a bit buggy?
08:41 orowith2os[m]: I'd like to correct that if it's at all wrong
08:41 jadahl: orowith2os[m]: Xorg either uses the modsetting driver (i.e. KMS) or vendor specific equivalent API in xorg video drivers
08:41 jadahl: "equivalent"
08:42 orowith2os[m]: I'm *assuming* in this case it would be the intel modesetting driver, but mblenc would have to confirm that
08:43 orowith2os[m]: not "intel modesetting driver", bleh, xf86-video-intel
08:44 mblenc: im on a radeon rx580, so have xf86-video-{amdgpu,ati}
08:53 MrCooper: orowith2os[m]: all major vendor Xorg drivers (with the notable exception of nvidia) have been using KMS since "forever" as well
08:53 MrCooper: what you're describing is outdated by over a decade
08:53 orowith2os[m]: good to know, thanks
08:54 orowith2os[m]: back to the main issue, mblenc: can you reproduce with the same adapter on another monitor, and without the adapter?
08:54 orowith2os[m]: worth a shot
08:55 mblenc: unfortunately, i do not have another monitor to reproduce this with. and my laptop doesn't have any vga-out
08:56 mblenc: actually, i can probably get a live-usb running on another laptop i have lying around
08:56 orowith2os[m]: alsooo what kernel version, and distro?
08:57 orowith2os[m]: worth changing those up a bit too
08:57 mblenc: 6.4.8-gentoo
08:57 mblenc: ill try with an ubuntu live-usb then
09:08 mblenc: i forgot how long it takes to download and prepare an ubuntu live disk
09:10 orowith2os[m]: Ventoy is a lifesaver, at least for hotswapping images :)
09:13 mblenc: i might very well have to set that up. i currently have like 7 usbs lying around, some for memtest, some for gparted, and the rest as insallation disks in various states of functioning
09:15 orowith2os[m]: they don't want you to know this but the Linux and debugging ISOs are all free, I have 12 of them on my USB with a full distro in a virtual hard disk too
09:50 mblenc: took a long time, but got the live-iso running
09:50 mblenc: ubuntu 23.0, works fine on both adapter and vga directly
09:51 orowith2os[m]: what kernel does Ubuntu ship?
09:51 mblenc: 6.2.0-20-generic
10:01 mblenc1: Testing my default laptop with the ubuntu live iso, maybe it is an issue with how i have configured my kernel and lsmod will give me some pointers
10:04 mblenc1: No, the same issue is present even on ubuntu
10:05 mblenc1: So the only constant is my laptop
10:05 orowith2os[m]: wanna go back a bit more, and try a 5.x kernel?
10:05 mblenc1: And even then, the hardware is probably at fault
10:05 orowith2os[m]: (Ubuntu 22.xx)
10:05 mblenc1: Sure, ill try that
10:08 mblenc1: The 22.xx iso is even larger (by 100MB) than the 23.xx version :(
10:16 mblenc: Installing the iso now, should just take another 30 minutes or so
10:45 mblenc: Trying ubuntu now, will take another 10 minutes to load and reach steady state
10:45 mblenc: I wish i had bought a faster pendrive :(
10:56 mblenc: The same issue occurs
10:57 mblenc: And nothing is shows in the dmesg output on ubuntu 22.04 either
10:57 orowith2os[m]: okay, good to know
10:58 orowith2os[m]: so it's probably not the kernel - could be hardware, or a long-lasting kernel bug :P
10:58 mblenc: I think its very probably a hardware issue
10:58 mblenc: As loathe as i am to say that
10:59 mblenc: I was able to run a test application under wayland and experienced the issue when the screen was filled with the rbg value of 220,220,220
10:59 mblenc: Or thereabouts
10:59 mblenc: Values of 200,200,200 ran just fine
10:59 mblenc: Which is *very* wierd to me
10:59 mblenc: Im wondering of its not a psu issue on the monitor side
11:00 mblenc: But that wouldnt explain why it works seemingly just fine under windows
11:00 mblenc: Or when connected to the other laptop
11:01 mblenc: Ive read (and personally experienced) that the rx 580 isnt particularly well supported under linux when it comes to stability and suspend
11:01 orowith2os[m]: my RX570 worked just fine fwiw
11:01 mblenc: But then again my laptop (GL702ZC) is a buggy platform in general
11:01 orowith2os[m]: I want to say it could be some difference in how everything's getting to your display on each OS; if Windows does something different, that could affect it
11:02 orowith2os[m]: anyways, from here I'll leave it to the more experienced folks
11:02 mblenc: Mine fails to start the display on boot sometimes, has random panics during boot, and suspend is completely broken
11:02 mblenc: But thats beside the point
11:16 kennylevinsen:hands daniels a cough drop
11:25 zumbi: Hello! On RISCV, we hit a segfault in a gtk4 testcase, which seems to be related to wayland, we have a backtrace - https://paste.debian.net/1288941/ - do you understand what could be happening there?
11:26 kennylevinsen: zumbi: maybe the surface is destroyed? maybe there's a bug in gtk4? Collecting wayland debug (WAYLAND_DEBUG=1) and looking at gdksruface-wayland.c would be useful
11:27 kennylevinsen: and printf debugging to the rescue
23:37 RAOF: 2
23:38 RAOF: Ahem.